Research map / Point clouds / Shape and surface completion
ProxyFormer: Proxy Alignment Assisted Point Cloud Completion With Missing Part Sensitive Transformer
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add the position extractor to the baseline (ablation) | ProxyFormer · PCN | CD-Avg: 11.08 -> 10.02 | Helped |
| Add the missing part sensitive transformer (ablation) | ProxyFormer · PCN | CD-Avg: 10.02 -> 7.74 | Helped |
| Add proxy alignment (ablation) | ProxyFormer · PCN | CD-Avg: 7.74 -> 6.77 | Helped |
| Use learned neighborhood position encoding instead of raw 3D coordinates (ablation) | ProxyFormer · PCN | CD-Avg: 9.63 -> 6.77 | Helped |
| Use learned neighborhood position encoding instead of MLP-style encoding (ablation) | ProxyFormer · PCN | CD-Avg: 7.83 -> 6.77 | Helped |
| Use 16 position neighbors rather than 8 (ablation) | ProxyFormer · PCN | CD-Avg: 6.86 -> 6.77 | Helped |
| Use 16 position neighbors rather than 32 (ablation) | ProxyFormer · PCN | CD-Avg: 6.92 -> 6.77 | Helped |
| Use Point Transformer features rather than MLP features (ablation) | ProxyFormer FAPE · PCN | CD-Avg: 8.08 -> 6.77 | Helped |
| Use Point Transformer features rather than DGCNN features (ablation) | ProxyFormer FAPE · PCN | CD-Avg: 7.18 -> 6.77 | Helped |
| Use Point Transformer features rather than scalar-attention transformer features (ablation) | ProxyFormer FAPE · PCN | CD-Avg: 7.53 -> 6.77 | Helped |
| Generate missing-part features instead of using random features (ablation) | ProxyFormer · PCN | CD-Avg: 7.90 -> 6.77 | Helped |
| Generate missing-part features instead of copying global features (ablation) | ProxyFormer · PCN | CD-Avg: 8.49 -> 6.77 | Helped |
Closest papers by meaning
- SymmCompletion: High-Fidelity and High-Consistency Point Cloud Completion with Symmetry Guidance (AAAI 2025 · similarity 0.88)
- DAPoinTr: Domain Adaptive Point Transformer for Point Cloud Completion (AAAI 2025 · similarity 0.87)
- CRA-PCN: Point Cloud Completion with Intra- and Inter-level Cross-Resolution Transformers (AAAI 2024 · similarity 0.86)
- PointCFormer: A Relation-Based Progressive Feature Extraction Network for Point Cloud Completion (AAAI 2025 · similarity 0.86)
- PCDreamer: Point Cloud Completion Through Multi-view Diffusion Priors (CVPR 2025 · similarity 0.85)
- Point Transformer V3: Simpler Faster Stronger (CVPR 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.